← Back to work CASE STUDY

Four people, running like eight.

The AI operating system behind a creative innovation team. Eighteen concurrent projects on one system, 70% less time on the weekly reporting, in daily use since February 2026.

A small creative innovation team was running roughly 18 concurrent projects with no shared system — knowledge stranded in folders and chat logs, AI used ad hoc with no shared patterns, and hours lost every week to manual reporting. The fix was an operating system — a context layer, reusable skills, and a knowledge base — that let four people operate like eight.

Every new project reinvented what the last one already figured out.

Knowledge lived in individuals, not systems. Eighteen projects with no portfolio visibility meant leadership asks took hours of manual synthesis. AI usage was ad hoc — no shared skills, no standards, no learning compounding across the team. Each project started from a cold folder.

COMPONENT 1

Context layer

Every conversation, decision, and client detail captured automatically into a shared, searchable memory — so no project starts from a cold folder.

COMPONENT 2

Reusable skills

Ten codified skills (research, synthesis, status, decisions, and more) plus four templates — so the team applies the same standards without re-explaining them every time.

COMPONENT 3

The operating hub

One registry for all eighteen projects, and commands that turn hours of weekly synthesis into a single step.

04 — THE RESULTS
70% less time

on weekly reporting that used to take hours

18 projects

running as one system

4 people

operating like 8

Performance evidence is captured continuously instead of reconstructed. A full cycle-end review that used to take a reconstruction marathon now happens in one sitting. The system gets smarter every week — each project adds to the registry, each learning reduces the next project's setup cost. In daily use since February 2026, and it costs cents a day to run.

Next case study

An AI Rescue: Busting Creative Director Bottlenecks

How a GPT-powered assistant reduced review cycles by 60% and freed the team to focus on strategy.